6. Parallel sets¶
6.1. 概要¶
6.2. Plotlyによる作図方法¶
6.3. MADB Labを用いた作図例¶
6.3.1. 下準備¶
import pandas as pd
import plotly.express as px
import warnings
warnings.filterwarnings('ignore')
# 前処理の結果，以下に分析対象ファイルが格納されていることを想定
PATH_DATA = '../../data/preprocess/out/magazines.csv'
# Jupyter Book用のPlotlyのrenderer
RENDERER = 'plotly_mimetype+notebook'
def add_years_to_df(df, unit_years=10):
"""unit_years単位で区切ったyears列を追加"""
df_new = df.copy()
df_new['years'] = \
pd.to_datetime(df['datePublished']).dt.year \
// unit_years * unit_years
df_new['years'] = df_new['years'].astype(str)
return df_new
def add_week_range_to_df(df, n_ranges=4):
"""weeksをn_ranges単位で分割"""
weeks = sorted(df['weeks'].unique())
def show_fig(fig):
"""Jupyter Bookでも表示可能なようRendererを指定"""
fig.update_layout(margin=dict(t=50, l=25, r=25, b=25))
fig.show(renderer=RENDERER)
df = pd.read_csv(PATH_DATA)
6.3.2. 雑誌別・年代別の合計作品数¶
# 10年単位で区切ったyearsを追加
df = add_years_to_df(df, 10)
df_plot = df.groupby(['mcname', 'years', 'cname'])\
['epname'].count().reset_index(name='weeks')
df_plot['single'] = df_plot['weeks']==1
df_plot['single_int'] = df_plot['single'].astype(int)
df_plot = df_plot.sort_values(
['mcname', 'years'], ascending=True)
fig = px.parallel_categories(
df_plot, dimensions=['years', 'mcname', 'single'],
color='single_int', color_continuous_scale=['Green', 'Red'],
labels={'years': '年代', 'mcname': '雑誌名', 'single': '一週のみ？'},
title='雑誌名・年代別の合計作品数')
fig.update_coloraxes(showscale=False)
show_fig(fig)